Henry M. Cowles


On his book The Scientific Method: An Evolution of Thinking from Darwin to Dewey

Cover Interview of July 07, 2021

The wide angle

Relevance can be an accident. When I started working on this project in 2010, Naomi Oreskes and Erik Conway had just published Merchants of Doubt, their account of organized science deniers who attack everything from climate data to vaccine schedules. Little did I (or even they) know how much worse things were going to get over the next decade.

By the time I sat down to write the book, we had entered the age of “alternative facts”; and when it was published four years later, both the pandemic and QAnon were in full-swing.

The distrust and polarization we’re seeing in 2021—with conspiracy theorists in Congress and politicians running on internet rumors, rather than running away from them—makes the tight-knit world of Merchants of Doubt seem almost quaint.

If relevance can be accidental, it can also be regrettable. Like many people, I wish we weren’t in the middle of a crisis of expertise—but here we are.

As a kind of “pre-history” of the scientific method, my book can’t tell you when the scientific method is being invoked properly or how better to wield it. Instead, the book ends with a question: what if pointing to a special method as the guarantor of scientific expertise is doing more harm than good?

The book hints at an answer by describing a different path. Most of the scientists I write about were interested in how science shades imperceptibly into other kinds of problem-solving. Their work didn’t lead to cynical conclusions about trust or truth, however. It simply insisted that science was a part of the world, an elaborate version of something that should be familiar to everyone.

Or at least, that was the idea.

In the end, as the book shows, that wasn’t the model of science and science’s publics that took shape in the twentieth century. For better and for worse, science came to seem like something that happened behind closed doors.

Distance has its benefits, especially when the process needs to be shielded from interference. But I argue that distance has its downsides, too. Most people don’t know how science actually gets done, which puts a premium on both media representations and the idea of a single method. What would happen if we closed the gap a little?